Eliminate Ringing at 10,000 mm Acceleration with Input shaping?

I'm in the process of tuning a 350mm Voron 2.4 for speed and quality. The printer is controlled by a RPI4, Fysetic Spider, Klipper and Fluid. In this video I show the results both before and after including the results of manual tuning Input Shaping. I resort to a ADXL345 accelerometer. Check to see the results of this adjustment. This will work for any coreXY printer and with slight modification will work with other types as well. Was it worth the effort?

An earlier version of this video was removed from this channel when helpful viewers let me know there were severe audio problems. It took a week to resolve due to a video editor bug, but this should be correct. Thank you too all who let me know!

All parts and documentation are linked to below.

These can be purchased at Amazon where I earn a commission that is directly applied to new content creation for this channel:

[tuningjohn]

Your video was extremely helpful for me. I also have a Voron 2.4 350x350mm and my settings turned out almost the exact same as yours. At first when built the cable to the ADXL345 I used a ribbon cable that I had lying around, and either the cable didn't work, or the accelerometer was bad. I switched to a CAT6 cable and another accelerometer and then it worked! Looking forward to your next upgrade and thanks for putting this online.

[andriusanryy]

Your video was extremely helpful for me :) thank you alot. Printing quality now is amazing on my voron 2.4 350. Reagards

[nicklong]

Glad I found your channel, top notch 👍

[valdolandar]

I would like to propose a `` non -orthodox 'solution for new users (I have not yet built the printer Voron 2.4, 350mm):
It would be about using the data you have achieved with the accelerometer and use them manually to implement it in the same Voron 2.4 350mm, so it is very possible that the data of the Input Shaper must actually be almost or equal.
shaper_type_x = mzv
shaper_feq_x = 50.4
shaper_type_y = mzv
shaper_feq_y = 36.2
I want to say that using those data that you use can already save us a lot of time and headaches. The same would be possible to share these data with different models of printers.

[jphalip]

Thanks for the great guide. Would you mind sharing the STL/3MF file you used to test the ringing? I assume it'd use multiple height modifiers to change the speed/acceleration through the print?

[slightwave]

Hi, I am also trying to figure out all this, and I still do not understand why people are so busy with 7000 accelerations if plastics such as PETG rarely support printing speeds greater than 90mm. and the flow limitation is almost reached. Well, you can get 150 there. But what is the point of such huge accelerations, it gives almost nothing in real time printing... what I don't understand? You can leave the acceleration at 4000 without all these complications and lose nothing.
Is it because I print PETG, is the situation different with other plastics?

[Rmilyard]

Did you change the max_accel: XXXX in printer.cfg? Mine is at 3000 right now.

[lcdconsultant]

Can you redo this with the CAN controller ADXL345 since you have that now?

[fluxcapacitor]

15:17: 7, 000 then 10, 000 mm/s² are high acceleration rate indeed! What were the print speeds set in the slicer for these two ringing-free prints?

[blackstoneriverworkshop]

Does it need to be on a CoreXY? I'm modding my ender 5 pro to use an Afterburner with the stock Cartesian system. Should I bother with this mod?

[fluxcapacitor]

If I understand correctly, this is not a permanent mount, is it? That is to say, you mount the ADXL onto the toolhead, connect it with the cable, make the resonance measurements, and finally unplug it and store it in a drawer, right? If so, how often should one mount and do the input shaping on a printer?

[BlownUpp]

Did you have a go at manually setting up the resonance testing? I’ve got a delta so the auto calibrate won’t work for me (no X and Y axis!), so I’m wondering how much difference it makes to probe multiple points and to go slower, or to adjust the accell_per_hz value (running at the default of 50 doesn’t make a sound - is it still effective even if it isn’t noticeably vibrating everything?)

[charlesliu]

Hi Kapman, nice video. Input shaping is definitely the way to go for FDM printer. I was inspired by this algorithm and went and developed my own version of input shaping with just a 32 bit control board. You can check out the video in here: